RE: is that a fork in the road?

["HUGHES,MARK (Non-HP-FtCollins,ex1)" <mark_hughes@non.hp.com>]

<digression>

>From: Bullard, Claude L (Len) [mailto:clbullar@ingr.com]
>>From: Tim Bray [mailto:tbray@textuality.com]
>>In fact (in particular from the Extreme Programming point of view), you
>>should arguably *always* use it as a model for software development.
>>And incrementalism applies well as an approach to lots of other
>>activities, too. -Tim
>As long as one doesn't have to deliver contractually 
>specified features (what most winning businesses do), that 
>works fine.   When one doesn't win, the VC capital runs out. 

  Ah, no.  I'm an XP programmer, and what XP promises is that the
customer will ALWAYS get whatever functionality is most important to
them, right away (well, at the end of an "iteration", which is 2 weeks
around here).  Waterfall development gives you a big-bang creation in
some longer period of time...  if it doesn't run late, if the design
specification is actually what the customer still wants by the end of
the development process (which is almost never the case), if the
engineers are so inhumanly smart and/or psychic that they can do perfect
design up front, if the requirements never change, and if absolutely
nothing goes wrong.  That any software at all is delivered by waterfall
is nothing short of a miracle.

  XP is primarily concerned with delivering "contractually specified
features".  It isn't perfect, but small, self-correcting steps are far
more accurate than a single aimed shot at a dodging target.

  Please take the time to actually observe the process, or even just
read the books, before making future inaccurate claims.

</digression>

-- <a href="http://kuoi.asui.uidaho.edu/~kamikaze/"> Mark Hughes </a>